Option 3 is the correct answer. Hierarchical firewall policies allow you to apply consistent firewall policies throughout your organization. Hierarchical firewall policies are containers for firewall rules, and by associating a policy with an organization or folder, all rules apply to the entire target organization or folder. Similar to Virtual Private Cloud (VPC) firewall rules, these policies include rules that allow you to explicitly deny or allow connections.

Option 1 is incorrect. You need to associate a hierarchical firewall policy with an organization or folder, not a hierarchical firewall rule.

Option 2 is incorrect. There is no feature called VPC firewall policy.

Option 4 is incorrect. VPC firewall rules apply to specific projects and networks. If you want to apply firewall rules to multiple VPC networks in your organization, take advantage of hierarchical firewall policies.

Option 4 is the correct answer. In Google Cloud, you can relax the limit by sending a limit relaxation request on the console. The easiest way to increase the project limit is to click the Manage Quotas link on the Create Project form in the console, enter a description about the need to increase the project limit, and submit your request.

Option 1 is incorrect. Instead of submitting an additional scope ticket to Google Cloud Support, make a cap relaxation request on the console.

Option 2 is incorrect. Technical Account Manager (TAM) is a service where you can get advice from experts. It helps keep your critical Google Cloud services in optimal condition so that you can operate effectively and efficiently in the cloud. It is used for more advanced technical support, not basic support for Google Cloud such as support for implementing cap relaxation.

Option 3 is incorrect. There are no contractual terms such as implementing additional contracts for Google Cloud Support. You simply need to make a normal cap relaxation request.

[Reference] https://cloud.google.com/resource-manager/docs/limits

Option 1 is the correct answer. By setting strong data integrity, you can ensure data integrity so that if there is data that has undergone a write operation or an update operation, the data before the update will not be read. However, with strong data consistency, I / O operations are more time consuming than eventual consistency models, as all duplicated copies must be synchronized before operations such as updates are completed. increase. This is a disadvantage.

Eventual consistency model is a consistency model that may cause a discrepancy between the data update process and the read process because the process is performed based on the loose consistency that the data should be synchronized as a result. It is not possible to prevent the data before the update from being read. On the other hand, you don't have to sync all the duplicated copies before operations such as updates are complete, which can speed things up.

Option 2 is incorrect. Even if you set strong data integrity, it can be read by multiple users at the same time. While matching the competition of each write process, the process is performed so that synchronization is successful.

Option 3 is incorrect. The network load remains the same, but the database processing load is higher, such as synchronizing all duplicated copies before operations such as updates are complete.

Option 4 is incorrect. There is no difference in price between eventual consistency and a strong data consistency model.

[Reference]

https://cloud.google.com/datastore/docs/articles/balancing-strong-and-eventual-consistency-with-google-cloud-datastore

Option 2 is the correct answer. You can set the traffic splitting for the entire project version by running the `gcloud app services set-traffic` command. Gcloud app services is a command to view and manage App Engine services. It is a command to set traffic division by setting gcloud app services set-traffic.

Option 1 is incorrect. gcloud app services browse opens the specified service in your browser.

Option 3 is incorrect. gcloud app services divide-traffic is not a valid command.

Option 4 is incorrect. gcloud services app set-traffic is not a valid command.

【reference】 https://cloud.google.com/sdk/gcloud/reference/app/services/set-traffic

Option 4 is the correct answer. The'kubectl run'command is used for configuration processes such as creating a POD after a Kubernetes cluster has been created. This allows you to run Docker images in your cluster. You can check the resources that can be created from the generator page of kubectl run.

Option 1 is incorrect. cloud run is not a command, but a service that develops and deploys scalable containerized applications on a fully managed serverless platform.

Option 2 is incorrect. kubectl container run is not a valid command.

Option 3 is incorrect. gcloud container operations is a command to get and list the operations of your Google Kubernetes Engine cluster.

[Reference] https://kubernetes.io/docs/reference/kubectl/conventions/#generators?

Option 2 is the correct answer. Use metric scopes to monitor your Kubernetes clusters. Metric scopes are a feature that can be used to store all configuration content for dashboards, alert policies, uptime checks, notification channels, and group definitions. You can use this to monitor your project.

Option 1 is incorrect. The sync is a function for executing the log distribution process. You can use a sync to forward some or all of your logs to a selectable destination.

Option 3 is incorrect. A pod is a single instance of a process running in a cluster. This is part of Kubernetes and is not used for monitoring.

Option 4 is incorrect. In preparation for monitoring a kubernetes cluster, it was necessary to set up a workspace in Cloud Monitoring, but now it's replaced by metric scopes.

【reference】

https://cloud.google.com/blog/products/management-tools/using-stackdriver-workspaces-help-manage-your-hybrid-and-multicloud-environment

Option 3 is the correct answer. The command line utility for manipulating Bigtable data structures is cbt. Bigtable is a fully managed, scalable NoSQL database service for large analytical and operational workloads with up to 99.999% availability. The cbt tool is then used as a command line interface for various operations in Cloud Bigtable. It is written in Go and uses the Go client library for Cloud Bigtable.

Option 1 is incorrect. gcloud is not used to interact with Bigtable's internal data structures.

Option 2 is incorrect. There is no command utility called bt.

Option 4 is incorrect. gsutil is a command for operating Cloud Storage.

【reference】 https://cloud.google.com/bigtable/docs/cbt-overview

Option 2 is the correct answer. You can configure an optional autoscaling policy to trigger the addition or removal of instances based on CPU utilization, Cloud Monitoring monitoring metrics, and load balancer load balancing capacity. Disk, network delay, and memory can trigger scaling when monitoring metrics for these resources are configured. Traffic distribution rate is not used as measurement.

[Index of target usage rate]

· Average CPU usage

· HTTP load balancing processing capacity based on either utilization or requests per second

· Cloud Monitoring indicators

Option 4 is the correct answer. Ruby is not available in Spark of Dataproc. Dataproc is a fully managed and scalable service for running over 30 open source tools and frameworks, including Apache Spark, Apache Flink, and Presto.

■ For Hadoop

The majority of developers use Java and Hive.

■ For Spark

You can use Java, Scala, Python, and R.

Therefore, options 1, 2 and 3 are the correct languages.

【reference】

https://cloud.google.com/dataproc/docs/concepts/overview

Option 1 is the correct answer. Take advantage of the App Engine version and the ability to split traffic to deploy the new version . Then, by dividing and routing the traffic from the current version to the new version, you can gradually increase the traffic volume to the new version. This will make switching to the new version at less of a risk during migration.

The traffic split feature allows you to specify a traffic distribution ratio to distribute traffic to multiple versions of the same service. Due to traffic splitting, A / B testing between versions allows you to control the pace at which features are deployed.

Option 2 is incorrect. The batch switching method can be significantly affected by problems with the new version or errors during migration.

Option 3 is incorrect. You should minimize the risk during migration by switching gradually rather than simply changing the routing destination.

Option 4 is incorrect. It is necessary to migrate by switching the routing, not by updating the previous version.